This article describes how to resolve 70% of issues related to Apache Cassandra that is bundled with Upsource and used for data storage.
- Cassandra gets restarted on a regular basis
- New commits are not getting indexed
- Some (all) revisions are stuck in "indexing in progress" state
Our statistics show that the most common reason for Cassandra failures is lack of memory assigned to it. Here are the steps on how to increase it:
- Stop Upsource with <upsource_home>/bin/upsource.sh stop
- Set new Xmx value with <upsource_home>/apps/cassandra/bin/cassandra.sh configure -J-Xmx5000m
- Start Upsource with <upsource_home>/bin/upsource.sh start
Note: We don't recommend setting Xmx value for more than 8 GBs.
To view current JVM options for a Cassandra, use the following command:
<upsource_home>/apps/cassandra/bin/cassandra.sh list jvm-options
If this didn't help, please submit a request with attached logs to the Upsource team.